powered by:
MagicWare, s.r.o.

On-line Fault Prediction Algorithm for the Pulse System

Authors:Kolokolov Yury, Department of Design and Technology of Electronic Systems, Russian Federation
Monovskaya Anna, Department of Design and Technology of Electronic Systems, Russian Federation
Hamzaoui Abdelaziz, CReSTIC, France
Topic:1.1 Modelling, Identification & Signal Processing
Session:Monitoring and Change Detection
Keywords: pulse systems, on-line dynamics identification, prediction, fractals

Abstract

In the paper the problems of the pulse system dynamics identification and its evolution prediction are analyzed. The offered approach to these problem solutions consists in one-to-one mapping of a system stable state by a point (a vector) within the special space designed, but a transitional process – by a vector trend. Then dynamics evolution forecasting is realized as the vector trend treatment relatively to the motion existence domains in this space. To illustrate the approach, the results of numerical modeling of the synchronous back voltage converter dynamics are used.